<会議発表論文>
User Interface of an Interactive Evolutionary Computation for Speech Processing
作成者 | |
---|---|
本文言語 | |
出版者 | |
発行日 | |
収録物名 | |
開始ページ | |
終了ページ | |
会議情報 | |
出版タイプ | |
アクセス権 | |
概要 | In this paper, we propose and evaluate three methods to improve interactive evolutionary computation (IEC) by decreasing user fatigue when an IEC displays sequential outputs for tasks such as music cr...eation, speech processing, or sound enhancement. The first method decreases human fatigue by using multiple sound sources. We evaluate two systems based on this method: a. system that assigns a different sound source to each individual and a system that assigns a different sound source to each generation. Subjective tests have shown that the first system is not as effective as the second system, which significantly decreases fatigue. The second method displays the previous best individual as a reference to compare subsequent individuals. Subjective tests have shown that this method significantly improves the IEC operability and has a tendency to improve the IEC convergence. The third method sequentially displays individuals tha.t cannot be spatially compared. Subjective tests have shown that this method significantly improves the IEC operability.続きを見る |
目次 | 1 Introduction 2 The Conditions of Experimental Evaluation 3 Evaluation of the IEC displaying multiple speech source 4 Evaluation of the method for displaying the elite 5 Evaluating the method for displaying the sound source in series 6 Discussion 7 Conclusion |
本文ファイル
ファイル | ファイルタイプ | サイズ | 閲覧回数 | 説明 |
---|---|---|---|---|
![]() |
2.89 MB | 115 |
詳細
レコードID | |
---|---|
主題 | |
登録日 | 2021.08.26 |
更新日 | 2021.08.27 |